<p>There are long-term uncertainties in precipitation estimation products, and the evaluation is the key to avoid or mitigate the impact of uncertainties, which is of great scientific significance to improve the monitoring of extreme climate events. This study aims at analyzing the accuracy performance of multi-source precipitation products, including the radar product (RADAR), satellite products (IMERG, GSMAP), the model product (ERA5) and merged product (CMPAS), under extreme weather and climate condition, such as rainstorm occurring on July 20, 2021 in Henan, a central region of China, from the perspectives of precipitation frequency-intensity distribution structure, diurnal variation and spatial-temporal evolution characteristics. All types of precipitation products are able to reflect the evolution of precipitation process, among which CMPAS performs the best, RADAR derives more hourly-scale heavy precipitation contributed to total precipitation, resulting in a serious overestimation, and the quantity of hourly weak precipitation samples derived from satellite and model products is overly high. The precipitation frequency of ERA5 is 40% higher than others. In terms of diurnal variation characteristics, CMPAS and RADAR do not show peak time phase shift and both reflect the asymmetry of the precipitation process (reaching the peak time quickly and then weakening slowly), while the satellite and model products have a significant peak time phase lag and reduce the asymmetry of precipitation process. CMPAS accurately reproduces the rapid increase in precipitation over the area in front of the mountain, RADAR overestimates precipitation in front of the mountain by a factor of one, and IMERG and GSMAP reflect a more easterly precipitation center. Although satellite and reanalysis precipitation products are not as effective as fused products in capturing extreme events, they still have potential application value in climate trend analysis.</p>